continuous_conv_transpose(filters, out_positions, out_importance, extents, offset, inp_positions, inp_features, inp_neighbors_index, inp_neighbors_importance_sum, inp_neighbors_row_splits, neighbors_index, neighbors_importance, neighbors_row_splits, output_type=tf.float32, align_corners=True, coordinate_mapping='ball_to_cube_radial', normalize=False, interpolation='linear', max_temp_mem_MB=64, debug=False, name=None)¶
Continuous tranpose convolution of two pointclouds.
A Tensor. Must be one of the following types: float32, float64, bfloat16.
The filter parameters. The shape of the filter is [depth, height, width, in_ch, out_ch]. The dimensions ‘depth’, ‘height’, ‘width’ define the spatial resolution of the filter. The spatial size of the filter is defined by the parameter ‘extents’.
A Tensor. Must be one of the following types: float32, float64.
A 1D tensor with the 3D point positions of each output point. The coordinates for each point is a vector with format [x,y,z].
out_importance – A Tensor. Must have the same type as filters. An optional scalar importance for each output point. The output features of each point will be multiplied with the corresponding value. The shape is [num input points]. Use a zero length Tensor to disable.
extents – A Tensor. Must have the same type as out_positions. The extent defines the spatial size of the filter for each input point. It is a 2D vector of the form [[x_size, y_size, z_size], ..]. For ‘ball to cube’ coordinate mappings the extent defines the bounding box of the ball. Broadcasting is supported for all axes. E.g. providing only the extent for a single point as well as only providing ‘x_size’ is valid.
offset – A Tensor. Must have the same type as out_positions. A 1D tensor which defines the offset in voxel units to shift the input points. Offsets will be ignored if align_corners is True.
inp_positions – A Tensor. Must have the same type as out_positions. A 2D tensor with the 3D point positions of each input point. The coordinates for each point is a vector with format [x,y,z].
inp_features – A Tensor. Must have the same type as filters. A 2D tensor which stores a feature vector for each input point.
A Tensor. Must be one of the following types: int32, int64.
The inp_neighbors_index stores a list of indices of neighbors for each input point as nested lists. The start and end of each list can be computed using ‘inp_neighbors_row_splits’.
A Tensor. Must have the same type as filters.
1D tensor of the same length as ‘inp_positions’ or zero length if neighbors_importance is empty. This is the sum of the values in ‘neighbors_importance’ for each input point.
inp_neighbors_row_splits – A Tensor of type int64. The exclusive prefix sum of the neighbor count for the input points including the total neighbor count as the last element. The size of this array is the number of input points + 1.
A Tensor. Must have the same type as inp_neighbors_index.
The neighbors_index stores a list of indices of neighbors for each output point as nested lists. The start and end of each list can be computed using ‘neighbors_row_splits’.
neighbors_importance – A Tensor. Must have the same type as filters. Tensor of the same shape as ‘neighbors_index’ with a scalar value that is used to scale the features of each neighbor.
neighbors_row_splits – A Tensor of type int64. The exclusive prefix sum of the neighbor count for the output points including the total neighbor count as the last element. The size of this array is the number of output points + 1.
output_type – An optional tf.DType from: tf.float32, tf.float64. Defaults to tf.float32. The type for the output.
align_corners – An optional bool. Defaults to True. If True the outer voxel centers of the filter grid are aligned with the boundary of the spatial shape.
An optional string from: “ball_to_cube_radial”, “ball_to_cube_volume_preserving”, “identity”. Defaults to “ball_to_cube_radial”.
Defines how the relative positions of the neighbors are mapped before computing filter indices. For all mappings relative coordinates will be scaled with the inverse extent, i.e. the extent becomes a unit cube. After that one of the following mappings will be applied:
’ball_to_cube_radial’: maps a unit ball to a unit cube by radial stretching. ‘ball_to_cube_volume_preserving’: maps a unit ball to a unit cube preserving the volume. ‘identity’: the identity mapping.
Use ‘ball_to_cube_radial’ for a spherical or ellipsoidal filter window and ‘identity’ for a rectangular filter window.
normalize – An optional bool. Defaults to False. If True the input feature values will be normalized using ‘inp_neighbors_importance_sum’.
An optional string from: “linear”, “linear_border”, “nearest_neighbor”. Defaults to “linear”.
If interpolation is ‘linear’ then each filter value lookup is a trilinear interpolation. If interpolation is ‘nearest_neighbor’ only the spatially closest value is considered. This makes the filter and therefore the convolution discontinuous.
max_temp_mem_MB – An optional int. Defaults to 64. Defines the maximum temporary memory in megabytes to be used for the GPU implementation. More memory means fewer kernel invocations. Note that the a minimum amount of temp memory will always be allocated even if this variable is set to 0.
debug – An optional bool. Defaults to False.
name – A name for the operation (optional).
A Tensor of type output_type. A Tensor with the output feature vectors for each output point.